SMART AGENT BASED SEARCH FOR ADMISSION IN INSTITUTIONS OF HIGHER LEARNING

International Journal on Integrating Technology in Education (IJITE) Vol.2, No.2, June 2013

DOI :10.5121/ijite.2013.2201 1

SMART AGENT BASED SEARCH FOR ADMISSION ININSTITUTIONS OF HIGHER LEARNING

Ricardo Adams1 and Suresh Sankaranarayanan2,3

1Department of Computing, University of WestIndies, Kingston, [email protected]

2Computing and Information Systems, Institut Teknologi Brunei, Brunei3Department of Computing, University of WestIndies, Kingston, Jamaica

[email protected]

ABSTRACT

Early admission systems saw people applying to universities by filling out applications forms and placingthem in suitable envelopes and sending them through the local postal agency. This was not considered tobe cost or time effective, and this method was also not efficient. This system however needed someimprovement due to the huge workload on administrators. So researchers and software developersimproved the system so that between 1999 and 2008 application and admission was done via the Internet.Also many Ranking system like ARWU, shanghai etc. been used for ranking the universities and collegesaround the world which would enable people choosing the universities and colleges for education onfactors like publication, funding, infrastructure and so.

The Internet has already brought the humans together in a new, exciting, and unexpected ways, and thesame is also happening to our prevalent adoption of digital mobile devices that has paved the way for thedevelopment of many innovative applications in the commercial domain. While considering such mobiledevices for an application towards higher education in an educational institution, there has been someamount of work done using intelligent agents. But still those agent based systems got some drawbackswhich motivated towards developing the present Agent based system to provide Smart agent based systemfor higher Learning search not in Jamaican context alone but also elsewhere with these drawbacksalleviated. The agents developed will be based on using fuzzy preference rules and heuristics, to makeaccurate decisions based on the user’s criteria or specifications using JADE-LEAP on Android handset.The system got Google map feature, intelligence in admission system and also warning for universitieswith low rating. These findings of this research will be presented as screenshots.

KEYWORDS

ARWU, JADE, LEAP, Android

1. INTRODUCTION

Early admission systems [1] saw people applying to universities by filling out applications formsand placing them in suitable envelopes and sending them through the local postal agency. Thiswas not considered to be cost or time effective, and this method was also not efficient.

The 1994 era saw Norwegian Universities and Colleges being the early pioneers of OnlineEducational Admission Systems (OEAS). This system however needed some improvement dueto the huge workload on administrators [2]. So researchers and software developers improved thesystem so that between 1999 and 2008 online application and admission was done via theInternet. The Internet has already brought the humans together in a new, exciting, andunexpected ways, and the same is also happening to our prevalent adoption of digital mobiledevices that has paved the way for the development of many innovative applications in thecommercial domain. The framework of M-commerce provides services like Mobile ticketing,Mobile banking, Mobile location based services, Mobile auctions, Mobile purchasing, etc. The

mailto:[email protected]

mailto:[email protected]


2

possibilities and opportunities for usage are endless when considering mobile devices, as auniversal device for mobile commerce applications. While considering such mobile devices foran application towards higher education in an educational institution, we generally come acrossselecting the best Universities with quality in overall teaching, available modern facilities andstudent resources, publications\research interest, student feedback, funding for courses, cleanenvironment and affordable rates. The tasks to be carried out towards getting these informationcan be humanly time consuming and sometimes also costly even with the using human agents.There of course exists lots of websites like webopedia, etc., which give the list of universities inthe country. But everyone would want to know how good the university is and what is the reviewlike by people. In addition there exists many ranking system like QS ranking, ARWU rankingand Times higher education [3] which ranks the universities in the country based on certainfactors like academic peer review, faculty student ratio, international faculty, internationalstudents, citation per faculty and so on. But to get access to this information the user has toperform enormous Google search. In fact to our knowledge there is no such system that existswherein ranking is based on feedback from the student community in addition to the one reportedby the university and also accessible from a mobile device. So based on these there has beensome research in terms of using Intelligent Agents in Mobile handset towards search ofuniversities which uses fuzzy preferences to search for institutions and thereby showing rating ofinstitutions based on feedback received from students compared to one received from thoseranking sites based on information received from Institutions. But still these agent based systemsgot some drawbacks in terms of location of university using Google map, intelligent admissionsystem by assessing the basic criteria and also sending warning to users and university authoritytowards lower ranking universities. These details are explained in forthcoming sections. Section2 talk on Literature survey on Artificial Intelligence in Higher Education System. Section 3 talkson Smart Agent based System architecture developed towards Institutions of higher learningfollowed by algorithm. Section 4 talks on implementation using JADE-LEAP. Section 5 talks onConclusion and Future work.

2. ARTIFICIAL INTELLIGENCE IN SEARCH AND ADMISSION

This section here gives a detailed literature survey on Artificial Intelligence used in searchsystems followed by their usage in Educational Systems

2.1. Artificial Intelligence in Search Systems

Agent based technology has been integrated in AI systems to do web searches and makedecisions for users. Agents are also utilized in expert systems and info bots. The concept ofsearch agents has become important in both mainstream computer science and ArtificialIntelligence (AI). Currently, search agents are being used in an increasingly wide variety ofapplications and they are the focus of intense relevance in many sub-fields of computer scienceand artificial intelligence [4]. Agent based systems and applications in the field of artificialintelligence, have given rise to new and improved software search agents which has evolved tomeet the users specifications. An agent is always viewed as something that acts in anenvironment. Search systems incorporating agents with search algorithms have been used tosolve various AI tasks. The key goal of intelligent search agents is to reduce user work andinformation burden. Intelligent search agents can provide services in online tutoring, searchingand filtering of information or data, and so on.

Since the Internet has become one of the most useful and powerful source of information, WebMining using web agents has become important which is the extraction of informationautomatically from web documents and services using data mining techniques. Inamdar et al2009 paper suggests getting rich source of information, from hyperlinks among pages or Webusage information using web data mining technology based on intelligent search engine. Overthe years AI has improved on various search algorithms such as best-first-search, depth-first-


3

search, breath-first-search and heuristic search to name a few. Modern day search systemsutilizes learning process components of AI search engines such as Google, Yahoo, Bing, andAsk.com to allow intelligent agents to communicate between users and network systems togather and formalize knowledge. Artificial intelligence has a wide area of applications withinsearch systems that are still under research. This research project will however focus on AI inEducational search and Admission Systems which are discussed in forthcoming sections

2.2. Artificial Intelligence in Higher Education Admission System

Gone were the days when applying to institutions required filling out blank forms and proofreading pages looking for errors. Afterwards packing the forms in 8 by 1 envelopes and sendingthem to Postal agencies [1]. We have entered a decade where admissions to higher learning cannow be done online. Most Universities now allow prospective students to apply to theirinstitutions via their websites. Some of these systems may differ from country to country andsometimes institution to institutions.

With the advent of Artificial intelligence, many admission systems have eliminated thetraditional admission process. Institutions are now offering Web-based applications with AItechnology integrated. This allows the student to apply without physically visiting the school tocollect or submit applications [5].

In one of the research reported [6] there was the need to implement a system to facilitate theadmission process. The system was web based and allowed the users to apply directly to theinstitutions. The system was called the Admissions Tutor. It was used to check all the requiredinformation for the student and it also ranked each students application. A student would beaccepted or rejected based on the rankings of the institution and the grades of the potentialstudent.

As the online Admission technology evolves, organizations like eadmissions.org allow parents toapply for placement of their children in nurseries and Kindergarten. This organization iscomprised of the 33 London Authorities and Surrey County Council and they work together tomake it easier for parents and care givers to apply for school places [7]

Admitting a student to a course is based on the individual university. They base their decisionson a number of factors, but primarily on the grades already achieved by the prospective students.As more applicants attain higher level of education, most institutions have devised a secondmeans of admission criteria. This includes results from GCSE or standard exams, the schoolsthey have attended and any personal information provided on the application. Most institutionshave implemented similarity detection software to determine if the personal statements werewritten by a third party and hence are able to reject a student application [2]. In terms of thisresearch, we will be rejecting the prospective student solely on not meeting the requirements forthe programs they have applied for.

Beyond the use of AI in web enabled systems, there has been research [8] which showed animplementation of an Agent-based Student Mobility Support system in mobile devices. Thisinvolved the development of an agent system that offered support to individuals wanting toparticipate in student exchange programs. The system allowed the user to search for foreigninstitutions and programs based on user defined criteria such as room availability, course taught,required course credits and teacher appointments and availability. The work didn’t have theability to compare ratings of the institutions, support feedback mechanism nor any intelligence toassist students when choosing an institution to apply.

As Online admission application becomes more and more popular and in recent times they havealso become more intelligent and allow for notifications.

Intelligent Agent Based University Search and Admission System [9] that was developedemploys an intelligent agent, which replicates a human agent towards performing the similar


4

search and admission activities that would incidentally improve the speed of the search andreduce cost considerably. The agents developed are based on using fuzzy preference rules andheuristics, to make accurate decisions based on the user’s criteria or specifications. The agenthere retrieves the list of universities based on user’s specification and displays it on to Mobilehandset of user along with feedback information of universities in terms of Quality of Education,Infrastructure, Environment, Cost etc. The feedback is based on what the students have reportedabout the university. The implementation was carried out using FIPA compliant JADE-LEAPagent development kit [10] on Android 2.2 handset.

Webtech a leading provider in online admission systems for Post Grad and Under Grad coursesis one such evolved system. Users are notified via their mobile devices and PDAs. Developers ofeducational systems have now integrated artificial intelligence in their system which is costefficient and has improve the user interaction experience when applying for institutions. Thismost times involves the use of Agents which will be discussed.

2.3. Intelligent Agent

The use of Agents is considered one of the most important aspects when improving on currentmethods for conceptualizing, designing and implementing software [10]. Agents are being usedin applications ranging from large critical systems such as air traffic control to relatively smallsystems such as email filters [4]. The ability of an agent to traverse around a network, allows foragent to agent interaction and communication of information. Mobile agent technology hasdeveloped into a number of research areas. With the use of algorithms, agents are able to makedecisions and return results for problem-solving tasks [11].

With the development of cutting edge toolkits and emulators, programmers can simulate mobileagent applications and hence enhance future agent technologies [12] . There are many agentToolkits available like Aglets, Concordia, JAFMAS and so. But for our work which involvesMobile, the Java Agent Development framework (JADE) is one which is suited that is a fullycompliant FIPA standards distributed middleware system with a flexible infrastructure allowingeasy extensions with add-ons such as LEAP for mobile phones [10]. As such, the proposedsystem will use JADE along with add-on LEAP to create a multi-agent solution that will supportAndroid 2.3 capable mobile devices

2.4. Ranking of Higher Education System

Rankings and ratings of colleges and universities are most times conducted by government,magazines, newspapers, and academics. An ordered combination of factors such as programs,departments, facilities etc. are used to assess a list of institutions in higher education [3]. Thenecessity for an international ranking of universities was highlighted in December 2003.Rankings of various types took into account aspects of excellence in research, student choicesand final success with demographics. Surveys were also use to assist in ranking evaluations.

Rankings evaluations on institutions are sometimes done within a single country or institutionsare assessed on a worldwide basis. The debate about ranking usefulness and accuracy is an ongoing one. The range in rating methodologies is ever expanding and the accompanying criticismof each indicates the lack of consensus in the field. There exists lot of websites like webopediawhich gives the list of universities in the country. But everyone wants to know how good theuniversity is and what is their review like by other people. Also there exists many ranking systemlike QS ranking, ARWU ranking and Times higher education which ranks the universities in thecountry based on certain factors like academic peer review, faculty student ratio, internationalfaculty, international students, citation per faculty and so on. But to get access to this informationthe user has to perform Google search. But there is no such system exists where ranking is basedon feedback from the student in addition to the one reported by the university and is accessiblefrom mobile device.


5

From the literature been surveyed it is cleared that there has been very little work reported onusing Intelligent Agents towards higher education admission system. Also there are quite amountof ranking methodologies used towards ranking of universities worldwide. But none of them usefeedback from students towards ranking but primarily depends on input from institutions only.So accordingly an Agent based system was developed on Android which enabled users toemploy intelligent agents in searching for institutions using fuzzy preferences based on criteriaand also retrieving institutions along with rating based on feedback from students. But still theagent based system got some drawbacks which was the main motivation for developing the smartagent based system which are explained in forthcoming section.

3. SMART AGENT BASED ARCHITECTURE FOR HIGHER LEARNINGSEARCH AND ADMISSION

It is important for us to explore and outline the process of what currently exists in the admissionsystems to a university, before we delve into details of our proposed system. The universityeducational institution admission system provides several sites and admission systems tostudents to enrol for admission to the university of their choice. Students may search throughyellow pages directory or walk-in and fill out application forms. These systems work buthowever they generally require the applicant to do the bulk of the searches and application.Students often find this to be a time consuming and costly exercise, and also frustrating at times.Sometimes the universities are not of the quality and accreditation that the applicant expected. Soseeing the inefficiencies in the current web based admission system, Intelligent Agent basedInstitutional Search and Admission system was developed in Android [9].

Intelligent Agent Based Institutional Search and Admission System [9] developed employs anintelligent agent, which replicates a human agent towards performing the institutional search andadmission activities that would incidentally improve the speed of the search and reduce costconsiderably. The agents developed are based on using fuzzy preference rules and heuristics, tomake accurate decisions based on the user’s criteria or specifications. The implementation wascarried out using FIPA compliant JADE-LEAP agent development kit on Android 2.2 handset.But we noticed some drawbacks in the system that has been the motivation for this improvedsmart Agent based system.

The first drawback was that the system does not use GMAP facility of Android handset towardsselecting the location for university search. Secondly the agent on retrieving the results to mobilehandset only displays the list of universities and their rating but not the location of universities.Thirdly the system only displays the rating of the university based on feedback received from thestudents and not as how popular the university is in terms of a particular program or disciplinebased on enrolment/graduation, research papers published, etc. based on past search experience.Also the earlier agent based system was unable to determine if the User met the requirements forthe program and hence it was unable to reject an admission made. The system onlyacknowledges for application made and do not possess intelligence towards admission system.Lastly the system possess not facility towards sending warning message to Institutions whoserating below certain value for improvement and also notify the user about those institutionsduring search process. These drawbacks were mitigated in the proposed Smart Agent basedsystem in this research which is explained below with architecture, Data Flow Diagram,Sequence diagram and algorithm

3.1 Smart Agent Based Architecture For Higher Learning Search And Admission

Smart Agent based architecture as shown in Figure 1 shows how agents communicates andtransfer information between themselves and how the agents interface with the user’s Androidphone. The smart agent models aspects of the admission process, by communicating and passinginformation between themselves and this is sent to the user’s mobile device. The Androidoperating system supports multiple concurrent communications between the smart agents,


6

allowing the agents to have the power of sending and receiving multiple messages at a timewhile still allowing full-duplex communication. The Android phone facilitates the use of threeagents implemented on the mobile phone handset as shown in Figure 1, giving the user a moreuser friendly experience when using the application. The Android technology has created muchpopularity since its introduction by Android Inc. a firm which was bought by Google in 2005.Users of the phone have the luxury and comfort of a PC running on their mobile device with alight weight platform that is flexible and facilitates data storage using SQLite. Android facilitatesa core set of applications such as, SMS program, calendar, email client, maps, browser, built-inGPS, and much more added features. A customized virtual machine named Dalvik allowing easyrunning of libraries inside the mobile device. The Android platform is considered to be a goodplatform for users and developers, due to the opening of much of its OS system. Because of thesereasons we decided to use the Android technology application.

Figure 1 Smart Agent Architecture

The Smart Agent Based Search and Admission System utilizes several components within itsarchitectural design. The main application platform requires an active mobile network inclusiveof internet access. The Smart Agents interact and communicate with each other to return resultsbased on the request of the user. The interaction from the user is to use the mobile graphical userinterface (GUI) to initiate a university search and admission process. The request iscommunicated by the agents throughout the system to gather some results to the user who thenmakes the confirmation. We will now explore the roles of each component within thearchitecture.

• User – This is the user who wants to acquire information about a given university. The user isresponsible for supplying the system the required search criteria that is used to generate the queryfor the requested data. The user will make an informed selection from the suggested resultspresented.


7

• Mobile Device – The preferred choice of mobile phone for the smart Agent was the Androidsmart phone with the 2.3 platform version. The research was primarily done on an AndroidEmulator and there were limitations faced due to security restrictions on the newer platform. Thisaffected some of commands issued within the application. The application was installed on themobile phone. Within the application, the GUI displays the search and admission engine.Possible search criteria are collected from the system and displayed to the user. The mobiledevice also hosts three intelligent agents for the system; these are the search, admission and thefeedback agents.

• Search Agent (SA) – The SA is hosted on the mobile device. The search agent is responsiblefor:

Collecting search criteria from the mobile device GUI.

Query the Central database Agent (CDBA).

Collect the generated results obtained from the CDBA.

Feed the results to the mobile device and display the result to the user

• Admission Agent (AA) - The AA is also hosted on the mobile device and has the responsibilityof performing all admission related operations with the system. The AA negotiates within thesystem. Its major functions are as follows:

Collect the admission details from the mobile GUI.

Perform pre-checks relating to the university availability and user criteria. This wouldhave been captured from the results from the Search Agent.

Perform checks on the applicant’s qualification for a desired program

Keep track of the user’s application to a program. This returns an application number tothe user.

• Central Database Agent (CDBA) – The CDBA is considered the main system due to it doingmajority of the work within the system. It is the link between the database and the resultsgenerated for the user to view. The CDBA has the responsibility to constantly perform multipleconcurrent communications between the Smart Agents, allowing the agents to collect andforward information. The CDBA is responsible for:

Use query generated from the SA and match the criteria with the information stored inthe database

Collect and store information about universities, with respect to fees, facilities, ratings,programs offered, location, contact details etc., from the respective university agents.

Collect and store the feedback results about universities from users and students in thecentral repository.

• University Agent (UA) - The UA resides inside each university. The UA communicates withthe CDBA and updates and provides information to the universities. The UA interfaces with theirrespective university system components to gather information about fees, facilities, ratings,degrees offered, discipline, location and information updated or stored on the respectiveuniversity’s website. It is important to note that information stored by the CBDA from the UA isused to provide query matching based on some user’s search criteria. The UA interfaces with theAA to provide universities with admission details, and this information is updated and stored indatabases to be accessed by the university’s administration.


8

• Feedback Agent (FA)– The FA allows the users to rate a university based on fees, educationand research facilities, infrastructure, security, general facilities, customer service andenvironment. The CDBA then stores these ratings for the FA to obtain and analyse the results offeedbacks over one, three and five year time periods. The overall ratings are fed back to thesearch agent to display to the user. The feedback agent is also responsible for notification toboth the mobile users and the Universities. When a University is below the recommended ratingswarning is sent in the form of an email.

3.2 Smart Agent Algorithm

Now based on Data Flow, Sequence diagram and Use Case diagram, we will now explore thealgorithm of the system based on the architecture and design. Presented below is the SmartAgent algorithm:

•User enters search details of a university from his/her mobile device and submit request for auniversity. Search criteria includes: Discipline, Country, fees, percentage price mark up,suggested facilities they want, etc.

•When submission is made the search agent passes request to the Central Database Agent(CDBA) that will communicate with respective university agents to search and locate a list ofadequate universities matching the search criteria of the user by applying the followingintelligence:

If a university is available for a lower fee range with exact or nearest matching offacilities in the same or another location.

If a university is available for the price range specified with exact or closest matching offacilities in the same or different location or country.

If a university is not available within the specified price range, it finds a suitableuniversity with the exact or nearest matching of facilities in the same or another location.

If no university is available within the user price range, agent finds a university with anyprice match with best matched facilities in same or different location.

•The university agents that satisfy the criteria will be displayed to the user’s (GUI) mobile devicefor a possible selection showing location of university, nearby places, driving direction etc usingGoogle Map.

•When a selection of a University is done, an Admission Agent (AA) is launched to facilitateadmission. This form is to be filled out by the user. When the User decides to submit, the systemwill verify if the applicant has met the requirement for the selected program. If not satisfied, theapplication is rejected.

•The Admission Agent then submits the form to the selected University for admission, thusallowing the university’s administrators to contact a prospective student.

•The system includes a feedback agent (FA) that allows the student to comment or give ratingson fees, programs, facilities and an overall rating of the university. Though this is present in theprevious system (Suresh and Cox, 2012), Smart Agent allows notifications to be sent via emailsand actions are taken when an institution falls below recommended ratings.

•The central database agent provides ratings on some aspect of the universities, which is thestored as averaged ratings calculated on all feedback received.

The implementation of Smart Agent based System been carried out using technologies like Java,JADE 4.0.1 with LEAP 4.0 and Android 1.0 add-ons, Android SDK and AVD Manager, Android2.3 emulator, MySQL and eclipse (IDE).


9

4. IMPLEMENTATION USING JADE-LEAP IN ANDROID

The implementation of the Intel-EISAS [9] was achieved using JADE 4.2 with LEAP andAndroid add-ons which is one of the latest Java Agent Development tool kit. When theapplication is launched for the first time, user is presented with the Login Screen as shown inFigure 2. If user doesn’t have an account, they will need to register as shown in Figure 3. Oncethe user has logged in as shown in Figure 4, user is redirected to the Home Screen of theApplication which allows performing the functionalities as shown in Figure 5.

Figure 2 Login Screen Figure 3 Register Screen

Figure 4 Login with Registered Credentials Figure 5 Home Screen for User

Once the user wants to search for an institution, user can initiate the Search Agent to find aninstitution. User would select criteria as given below that would influence their institution ofchoice which are:

•Discipline: This would include major fields of study such as Medicine, Information Technologyand Business


10

•Preferred mode of study. Whether the applicant is interested in Face to Face Program online andor a Part Time program

•Expected Cost. This is cost range for which the User can select how much money they arewilling to pay to enroll in a program

•Country for the school. If the User is willing to travel or interested in a program offered in acountry they would select from a drop down menu.

•Scholarship: Whether or not they want scholarship funding program.

• Distance: Proximity to University from place of residence

The User has the option of running the search on the provided selections or to continue thesearch with the inclusion of required facilities that the institution would have to provide. Thefacilities include a Library, Tutorials, Reading Room, Transportation Clubs and societies, Hallsof Residence , Gym, Sports, Laboratory in the event the student wants to live on the campus.Based on criteria selected, search agent does the search of institution and associated programme.Once an institution is found, the user will get the option of applying or get rating information onthe institution. In situations where an institution is not found based on the user’s criteria, anadvance search is done. These are explained below showing various agent intelligence scenarios.

Let us consider a scenario where the user chooses an Online IT Program located in Jamaica for acost ranging $5000 to $7999 with no markup and other criteria like scholarship funded, distanceto university and so forth as shown in Figure 6. The User has the option of running the search onthe provided selections or to continue the search with the inclusion of required facilities that theinstitution would have to provide towards search like Library, Sports, Halls of residence asshown in Figure 7. Now based on the criteria and facilities included, search agent contacts thecentral database agents which are repository of all institutions with facilities. The agent possessthe intelligence to search for institutions matching criteria with price of 5000-7999 in Jamaicaand finds one institution which is University of Technology Jamaica with cost of 6500 alongwith rating for Programme and university as shown in Figure 8.

Let us consider another scenario where User searches for a Face to Face IT Program located inJamaica for a cost ranging $8000 to $10,999 with university nearby and other facilities likeLibrary, sports, Halls of Residence similar to one shown in Figure 6 and 7. Once user selectedthe criteria, the search agent is initiated which contacts the Central database agent towardssearching for institutions with price range of 8000 to 10,999 in Jamaica for Face to Face ITprogram. But the search agent could not retrieve institutions matching the all the user’s criteria inJamaica instead the search agent applies the intelligence and retrieves the institutions to the userswith best matching user’s criteria with facilities in current and neighboring location as shown inFigure 9. The user can now make a decision whether to go with suggested institution or not.


11

Figure 6 Search Agent user criteria Figure 7 Facilities Inclusion

Figure 8 Search Result Figure 9 Best Match and Neighboring Institutions

Let us consider another scenario where User searches for a Online IT Program located in Jamaicafor a cost ranging $8000 to $10,999 with university nearby and other facilities like Library,sports, Gym, Laboratory, Reading Room similar to one shown in Figure 6 and 7. Once userselected the criteria, the search agent is initiated which contacts the Central database agenttowards searching for institutions with price range of 8000 to 10,999 in Jamaica for Face to FaceIT program. But the search agent could not retrieve institutions matching the user’s criteriainstead the search agent applies the intelligence and retrieves the best approximated Programbased on the cost chosen and facilities chosen. It also returns a suggested institution institutionsto the users as shown in Figure 10. The user can now make a decision whether to go withsuggested institution or not.

Let us consider another scenario where User searches for Face to Face IT Program offered inCanada for any cost with university nearby and other facilities like Library, Transportation, andHalls of residence similar to one shown in Figures 6 and 7. Once user selected the criteria, the


12

search agent is initiated which contacts the Central database agent towards searching forinstitutions with any price in Canada for Face to Face IT program. The Search Agent returnsInstitutions offering a Full Time IT Program located in Canada at any given cost as shown inFigure 11.

Let us consider final scenario where User searches for Online IT Program offered in UnitedStates with cost ranging from $17000 to $19999 with university nearby and other facilities likeTutorials, Laboratory, Clubs and Societies, Gym similar to one shown in Figures 6 and 7. Onceuser selected the criteria, the search agent is initiated which contacts the Central database agenttowards searching for institutions with price of 17000 to 19999 in United States for Online ITprogram. The search agent could not retrieve institutions matching the user’s criteria instead thesearch agent applies the intelligence and retrieves the institutions the best approximated Programbased on the cost chosen and facilities chosen in neighboring location. It also returns suggestedinstitutions to the users as shown in Figure 12. The user can now make a decision whether to gowith suggested institution or not.

Figure 10 Best Match and Neighboring Institutions Figure 11 Search Results


13

Figure12 Best Match with Neighboring Institutions Figure 13 Check Ratings

Figure 13 shows the Universities and Programs for which the user can check the ratings. Once anInstitution is chosen as shown in Figure 19 , the smart agent based system will automaticallypopulate Programs for the chosen Institution as shown in Figure 14. The User can click “GetRatings” under the Institution to view the ratings for the Institution and all available facilities asshown in Figure 15. In addition to viewing the rating of Institutions and selected facilities, theagent based system also allows user to select Programme offered in university which will displaythe ratings for the chosen program as shown in Figure 16 and 17.

Figure 14 Institution Selection Figure 15 Program Selections


14

Figure 16 Ratings for Facilities - Institution Figure 17 Rating for Programme

When user is satisfied with a program offered at an institution based on rating values, they canbegin the application process. This involves checking if they meet the requirements for theprogram chosen by entering their qualifications by invoking Admission Agent as shown inFigure 18. User shows the list of institutions and Programme offered in the respective institutionto be selected for applying. Each time a program is chosen, the prerequisites for the program isshown as seen in Figure 18. As seen for BA in History, the prerequisites are different than a BScin Computer Science as shown in Figure 19.

Figure 18 Admission Agent Figure 19 Programme Prerequisite

Once program is selected, the use selects “Enter Qualification”. The User will enterQualifications which consist of the institutions they went to and qualifications gained. The periodof attendance and qualifications earned. For each program selected, there are prerequisites thatthe applicant has to meet. The user is given the opportunity to enter more than one qualification,


15

thereby allowing them to further show their qualification for the course. If the similarqualifications are entered, the system will produce a notification. The Admission agent herepossesses intelligence to check applications that they are met before admitting the person to theprogram. If not, they are automatically rejected. This is an improvement over previous systemwhich only acknowledges once applied irrespective requirements met or not. These are shown inFigures 20 and 21. Also if the applicant enters the same subject twice, warning is issued byAdmission agent. Figure 31 shows where the user enters CAPE qualification. Applicant will alsomeet the requirement since he has Math, Information Technology and 5 CXCs. Figure 22 showswhere Admission Agent possess intelligence towards accepting Application based onrequirements and accordingly application code is provided for future reference as shown inFigure 23.

Figure 20 Qualification & subjects Figure 21 Application Rejected

Figure 22 Satisfying Requirements Figure 23 Successful Application

In addition to applying for institution of choice, the user can also get the location of theinstitution from the Google Map feature. In addition to getting location of Institution, user canalso determine the best way to get the school and also view other locations near to the institutionof choice. The User also has the ability to view all places nearby the institution. This could beadded reason to attend the school. When an institution is selected, the user can also get directionwhether by Walking, Driving or Commuting based on their location. These are shown in Figures24 to 26.


16

Figure 24 Institution Selection- Google Map Figure 25 Location of Institution

Figure 26 Driving Directions Figure 27 Notification on Institution

Figure 28 Notification on selected program Figure 29 Institution is made inactive


17

Based on the result Search agent, User is able to click on Notification for the university selectedto see if there is any alert for them to view. The notifications that are generated will indicate ifeither the institution or program selected is rated below the minimum requirements. The alertwould indicate the possibility of the institution no longer existing which are shown in Figures 27and 28.

Under the admin profile, the user can send a warning letter to the institution when they fall belowthe minimum ratings of 2. When a warning is to be produced, the smart agent calls the in-builtEmail App which allows for the emailing of the warning letter. The warning letter details are alsosaved to the database for future reference. If rating does not improve over time, Institution in thiscase is University of Technology is made inactive and will longer be available to future students.These are shown in Figure 29. Now say user logs into the system to check the rating ofUniversity of Technology which is no longer available as it is made inactive. Now after seriousconsideration and improvement by university, University council activates the university and isavailable. Once made active, user is now able to select university for applying and viewing therating

5. CONCLUSION & FUTURE WORK

It is evident in today’s world that there is a vast influx and birth of University institutions thathas flooded the internet or the World Wide Web (WWW), and many of them lacks accreditationand credibility. But with mobile agent technology these vulnerabilities can be adversely andgreatly reduced.

Mobile Search Agents can be used to verify and validate an accredited university within theapplicant’s geographical location. Once a university name is obtained, the agent automaticallysources the information from accredited government and regulatory bodies’ databases andwebsites in that countries locality to ascertain the validity and accreditation of the nameduniversity body. Ranking systems for world universities such as Shanghai Rankings, QS WorldUniversity Rankings, Times Higher Education Reputation Rankings and Webometrics, are notranking systems facilitating feedback ratings of students based on a university’s educationquality, research facilities, fees, environment, customer service and security. So accordinglyAgent based system for Institutional search and Admission was developed (Suresh and Cox,2012) i.e Intel-EISAS that mitigates above mentioned shortcomings. The system surroundsaround Jamaican university and college institutions towards search based on user’s criteria byapplying fuzzy preferences. Though the above mentioned Agent based system saves time andefficient in selecting universities, but still it got some drawbacks which are mitigated in currentsmart agent based system developed in this research project. The results and tasks are generatedby autonomous agents designed in a way that allow for cooperation, interoperability andcommunication within the system enabled the coordination of the entire search, selection andadmission process thus allowing little or no human intervention resulting in a more time and costeffective process. In future the system developed can be integrated with agent learning capabilitywhich allows the agents to use past information to assist in searching for institutions rather thanasking users to select all criteria. Also there is no payment feature as this requires Securityimprovement. This requires added WAP services in relation to credit card payments with thesystem.

REFERENCES1. Guernsey, L., 1999. Spam Your Way to a Good Education; Online Application Forms Add to

College Admissions Frenzy. Available at: http://www.nytimes.com/1999/12/23/technology/spam-your-way-good-education-online-application-forms-add-college-admissions.html?src=pmhttp://foba.lakeheadu.ca/serenko/Agent_Toolkits_Working_Paper.pdf;accessed April 30, 2013.

http://www.nytimes.com/1999/12/23/technology/spam-

http://foba.lakeheadu.ca/serenko/Agent_Toolkits_Working_Paper.pdf


18

2. Wikipedia 2010. Norwegian University and Colleges Admission service. Available at:http://en.wikipedia.org/wiki/Norwegian_Universities_and_Colleges_Admission_Service;accessed April 28, 2013.

3. College and university rankings (2010). Available at:http://en.wikipedia.org/wiki/College_and_university_rankings; accessed April 30, 2013.

4. Zambonelli, F., Jennings, N. Wooldrigde, M. 2003. Developing Multi-agent Systems: The GaiaMethodology. Transactions on Software Engineering and Methodology. Available at:http://www.csc.liv.ac.uk/~mjw/pubs/tosem2003.pdf; accessed April 30, 2013.

5. Naresh, S. n.d. Benefits of Online Schools Admission System. Available at:http://admission.ezinemark.com/benefits-of-online-schools-admission-system-31dfa34e691.html;accessed April 30, 2013.

6. Vignesh, B. 1993. Software Requirements Specification for Online University AdmissionSystem. VIT University 1993, IEEE Guide to Software Requirements Specifications (Std 830-1993)

7. EAdmission n.d. An online Admission System for parents to find schools for their children. Thiscan be found at: https://www.eadmissions.org.uk/eAdmissions/app?; accessed April 30, 2013.

8. Ganzha, M., Kuranowski, W., Paprzycki, M., Rahimi, S. and Szymczak, M. 2005. Designing anAgent-based Student Mobility Support System. Department of Computer Science, GizyckoPrivate Higher Educational Institute, Daszynskiego 9, 11-500.

9. Suresh, S. and Cox, L. 2012. Intelligent Agent based University Search and Admission System -Android Environment. International Journal of Computer Information System and IndustrialManagement Applications, 4: 35-42.

10. Bellifemine, F., Caire, G., & Greenwood, D. 2007. Developing Multi- Agent Systems withJADE. London: Wiley Publishing.

11. Reilly, D. 1998. Mobile Agents - Process migration and its implications. Available at:http://www.davidreilly.com/topics/software_agents/mobile_agents/; accessed April 23, 2013.

12. Yang, J., and Elhadi, S. 2010. Performance Evaluation of Agent Toolkits. Computer ScienceDepartment Acadia University Nova Scotia, Canada B4P2R6. Available at:http://www.springerlink.com/content/0a4nqqqbxfmblxx5/fulltext.pdf; accessed April 28, 2013.

Authors

Ricardo Adams holds Bachelor’s degree in Computer Science and Electronics in 2008 and currentlypursuing Master’s degree in Computer Science from University of WestIndies, Jamaica since 2011. He iscurrently working as an IT Audit Consultant and also Part Time Lecturer in University of TechnologyJamaica too. He is well versed in programming languages like Java, Php and so forth. His current researchinterests are mainly in Intelligent Agents, Mobile Computing.

Prof.(Dr) Suresh. Sankaranarayanan is currently an Associate Professor, Department of Computer &Information Systems, Institute of Technology, Brunei (ITB – A technological university) and also VisitingProfessor , Department of Computing, University of WestIndies, Kingston, Jamaica. He has supervisedaround 30 research students leading to M.Sc, ME, M.Phil and M.S degrees and currently supervisingstudents leading to M.sc, M.phil and Ph.d respectively. He has got to his credit, as on date, about 50 fullyrefereed research papers published in the Proceedings of major IEEE international conferences, as BookChapters and in International Journals. His current research interests are mainly towards ‘Mobile andUbiquitous Computing - Wireless Sensor Networks, Mobile Commerce, Intelligent Agents’ used in theHealth, Commercial and Engineering sectors.

http://en.wikipedia.org/wiki/Norwegian_Universities_and_Colleges_Admission_Service

http://en.wikipedia.org/wiki/College_and_university_rankings

http://www.csc.liv.ac.uk/~mjw/pubs/tosem2003.pdf

http://admission.ezinemark.com/benefits-of-online-schools-admission-system-31dfa34e691.html

www.eadmissions.org.uk/eAdmissions/app

http://www.davidreilly.com/topics/software_agents/mobile_agents/

http://www.springerlink.com/content/0a4nqqqbxfmblxx5/fulltext.pdf